JP2012517291A - Calculation of cardiovascular parameters - Google Patents

Abstract

A method is described for measuring a patient's cardiovascular parameters regardless of whether the patient is in a normal or abnormal hemodynamic state. These methods include determining whether the patient is in a normal or abnormal hemodynamic state and then applying an appropriate model to the patient data to determine cardiovascular parameters for the patient. A multivariate Boolean model is used to determine whether the patient is in a normal or abnormal hemodynamic state, and then a multivariate statistical model is used to calculate appropriate cardiovascular parameters. By having accurate cardiovascular parameters for patients in abnormal hemodynamic conditions, accurate calculations can be obtained for treatment related parameters such as cardiac output and stroke volume, for example.

Description

(Cross-reference of related applications)
This application claims priority to US Provisional Application No. 61 / 150,991, filed Feb. 9, 2009, entitled "Calculating Cardiovascular Parameters" as assigned to the assignee. , The entire contents of which are incorporated herein by reference.

  Stroke volume (SV), cardiac output (CO), end-diastolic volume, ejection fraction (EF), stroke volume fluctuation (SVV), pulse pressure fluctuation (PPV), and systolic blood pressure fluctuation Indicators such as (SPV) are important not only for disease diagnosis, but also for "real time", i.e. continuous monitoring of patient clinical significant changes. For example, healthcare workers are also interested in preload dependence, water responsiveness, or volume responsiveness, and cardiac output in both human and animal patients. Thus, most hospitals have some form of equipment that monitors one or more indicators related to the heart to alert one or more indicated changes that occur to the patient. Many techniques are utilized, including invasive techniques, non-invasive techniques, and combinations thereof, and many have been proposed in the literature.

  A method for calculating a patient's cardiovascular parameters is described. These methods include providing the patient's arterial pressure waveform data and analyzing the arterial pressure waveform data to determine whether the patient is in an abnormal state. If it is determined that the patient is in an abnormal state, then a multivariate statistical model is applied to the arterial pressure waveform data to determine the patient's cardiovascular parameters. This multivariate statistical model can be created, for example, using data from a patient in an abnormal state. If it is determined that the patient is not in an abnormal state (ie, in a normal state opposite to the abnormal state), then a multivariate statistical model is applied to the arterial pressure waveform data to determine the patient's cardiovascular parameters. decide. This multivariate statistical model can be created, for example, using data from patients who were not in an abnormal state (ie, patients who were in a normal state opposite to the abnormal state).

  In addition, methods for measuring cardiovascular parameters of patients with high cardiac output and patients with non-high cardiac output are described. These methods include providing the patient's arterial pressure waveform data and analyzing the arterial pressure waveform data to determine whether the patient's peripheral arterial pressure has been separated from the central aortic pressure. Analysis of arterial pressure waveform data to determine whether the patient's peripheral arterial pressure has been separated from the central aortic pressure is, for example, data from both patients who have undergone separation and those in normal hemodynamic status Can be achieved using a multivariate Boolean model generated based on If it is determined that the patient's peripheral arterial pressure has been separated from the central aortic pressure, then the arterial pressure waveform data is analyzed to determine if the patient has high cardiac output. Analysis of arterial pressure waveform data to determine whether a patient has high cardiac output can include, for example, data from both high cardiac output patients and patients in normal hemodynamic conditions. It can be achieved using a multivariate Boolean model generated based on it. If the patient is determined to have high cardiac output, then a multivariate statistical model is applied to the arterial pressure waveform data to determine the patient's high cardiac output and isolated cardiovascular parameters To do. This multivariate statistical model can be created, for example, utilizing data from patients experiencing high cardiac output status and separation. However, if it is not determined that the patient's peripheral arterial pressure is separated from the central aortic pressure, or if the patient is not determined to have high cardiac output, then a multivariate statistical model is applied to the arterial pressure waveform data. Thus, the patient's normal cardiovascular parameters are determined. This multivariate statistical model can be created, for example, utilizing data from patients who have not experienced segregation or high cardiac output status (ie, patients in normal hemodynamic status).

It is a figure which shows the pressure waveform simultaneously recorded in the ascending aorta (aorta), the femoral artery (thigh), and the radial artery (radius) of the pig model in a normal hemodynamic state. Simultaneously in the ascending aorta (aorta), femoral artery (femoral), and radial artery (radius) of a porcine model during endotoxic shock (septic shock) resuscitated with a large volume of fluid and pressor It is a figure which shows the recorded pressure waveform. The logic of the method described herein for deciding whether to apply a normal multivariate statistical model to determine cardiovascular or to use a multivariate statistical model based on abnormal conditions It is a block diagram which shows an example. Logic in the methods described herein for determining whether to apply a normal multivariate statistical model to determine arterial tone, or to use a high cardiac output multivariate statistical model It is a block diagram which shows an example. It is a figure which shows an example of the composite blood pressure curve over the heart cycle which is one heartbeat interval. It is a figure which shows the discrete time expression of the pressure waveform of FIG. It is a figure which shows the area | region based on the cardiac contraction part in an arterial pressure waveform. FIG. 6 is a diagram showing a statistical distribution of regions based on cardiac contraction stages in arterial pressure waveforms of normal patients and patients with high cardiac output. It is a figure which shows the cardiac contraction period in an arterial pressure waveform. FIG. 6 is a diagram showing a statistical distribution of cardiac contraction periods in arterial pressure waveforms of normal patients and patients with high cardiac output. It is a figure which shows the cardiac contraction period and cardiac expansion period in an arterial pressure waveform. Statistical distribution of periods in the diastole phase of patients with high heart rate are shown by normal hemodynamic status (dashed line) and high cardiac output status (thick line), and further distribution in all related patients (thin line) FIG. Statistical distribution of time during the systole stage of patients with high heart rate is shown by normal hemodynamic status (dashed line) and high cardiac output status (thick line), and further distribution for all related patients (thin line) FIG. FIG. 2 is a block diagram illustrating an example of a system for performing the methods described herein.

  A method is described for measuring cardiovascular parameters, such as a patient's arterial tone, whether or not the patient is in a normal or abnormal hemodynamic state. These methods include determining whether the patient is in a normal or abnormal hemodynamic state and then applying an appropriate model to the patient data to determine cardiovascular parameters for the patient. The method described herein produces a separate multivariate statistical model for an abnormal hemodynamic patient compared to a normal hemodynamic patient because the patient is once abnormal. This is because once the hemodynamic state is reached, a model based on a patient in a normal hemodynamic state is not accurate. Having an accurate cardiovascular parameter value for a patient in an abnormal hemodynamic state, for example, an accurate calculated value can be obtained using parameters based on the cardiovascular parameter in question. This allows the clinician to perform appropriate treatment on the patient.

  The methods described herein are illustratively used to determine a patient's arterial tone, regardless of whether the patient is in a normal hemodynamic state or a high cardiac output hemodynamic state. obtain. As used herein, the term “arterial tone” refers to the combined result in vascular compliance (ie, greater vascular elastic compliance) and peripheral resistance (ie, small peripheral vascular flow resistance). Similarly, as used herein, the phrase “high cardiac output and vasodilatation” means the condition where peripheral arterial pressure and its flow are separated from central aortic pressure and its flow, and the term “peripheral “Artery” is intended to mean an artery located away from the heart, such as the rib, thigh, or brachial artery. “Isolated arterial pressure” means that the normal relationship between peripheral arterial pressure and central pressure (ie, arterial pressure is amplified towards the periphery) is invalid. This further includes situations where the peripheral arterial pressure is not proportional to or affected by the central aortic pressure. Under normal hemodynamic conditions, blood pressure readings that increase with distance from the heart are obtained. Such an increase in blood pressure is shown in FIG. That is, the amplitude of the pressure wave measured at the radial artery is greater than the pressure measured at the femoral artery, and even greater than the aortic pressure. These pressure differences are related to wave reflection, which increases the pressure towards the periphery.

  This normal hemodynamic pressure relationship, that is, the pressure that increases with distance from the heart, is often relied upon in medical diagnosis. However, under high cardiac output / vasodilation conditions, this relationship may be reversed and the arterial pressure is less than the central aortic pressure. This reversal is due, for example, to arterial tone in the peripheral blood vessels, which is thought to affect the wave reflection described above. Such a high cardiac output state is shown in FIG. That is, the amplitude of the pressure wave measured at the radial artery is smaller than the pressure measured as the femoral artery, and even smaller than the aortic pressure. Drugs that dilate small peripheral arteries (such as nitrates, angiotensin converting enzyme inhibitors, and calcium inhibitors) are thought to contribute to the high cardiac output status. These types of severe vasodilation are often observed in situations immediately after cardiopulmonary bypass (coronary artery bypass) where the radial artery pressure is not affected by the pressure in the aorta. Substantial pressure differences between the center and periphery, where peripheral arterial pressure is unaffected by central aortic pressure, can be severe in patients with severe sepsis who are treated with high volumes of fluid and high doses of vasopressors leading to severe vasodilation. Always observed. A very similar condition is observed in patients with end-stage liver disease. As will be appreciated by those skilled in the art, certain treatments for patients in a normal hemodynamic state are different from those in a patient in a high cardiac output state. Therefore, the disclosed method detects the patient's vascular condition and utilizes appropriate calculation methods as needed to determine the patient's exact cardiovascular parameters such as arterial tone and cardiac output.

  A method for measuring cardiovascular parameters of a patient in a normal or abnormal hemodynamic state as described herein typically comprises providing the patient's arterial pressure waveform data and then analyzing the data (As used herein, the term “arterial pressure waveform data” means any other signal that is proportional to, derived from, or a function of the arterial pressure waveform data or the arterial pressure signal. Intended to). Depending on one or more results of each analysis step, the next analysis to be performed is determined. An example of the method steps is shown in FIG. First, the arterial pressure waveform is analyzed to determine if it is in an abnormal state (denoted by 10). If it is determined that the patient is in an abnormal hemodynamic state, then a calculation using the arterial pressure waveform data is performed to determine the cardiovascular parameters in the abnormal state (shown at 20). If it is determined that the patient is in a normal hemodynamic state, then a calculation on the arterial pressure waveform is performed to determine cardiovascular parameters in the normal hemodynamic state (denoted 30). In addition, other values associated with cardiovascular parameters, such as arterial tone, can be calculated (shown at 40) using appropriate values as needed.

A further example of a particular method step described herein is shown in FIG. The method includes measuring arterial tone between a patient in a high cardiac output hemodynamic state and a patient in a normal hemodynamic state. In this method, the arterial pressure waveform is first analyzed to determine if the patient's peripheral arterial pressure has been separated from the central aortic pressure (shown at 10). If it is determined that the patient's peripheral arterial pressure has been separated from the central aortic pressure, then the arterial pressure waveform data is analyzed to determine whether the patient has high cardiac output (shown at 20). ). If the patient is determined to have higher cardiac output, then a multivariate statistical model is applied to the patient's arterial pressure waveform data to determine the patient's high cardiac output arterial tone ( X _decoupled ) (denoted 30). If the patient's peripheral arterial pressure is not determined to be separated from the central aortic pressure, or the patient is not determined to have high cardiac output, then a multivariate statistical model is applied to the patient's arterial pressure waveform data. The normal arterial tone of the patient is determined (X _normal ) (indicated by 40). Next, the arterial tone (ie, C _normal or C _decoupled ) appropriate for those conditions of the patient is used to calculate the correct stroke volume, cardiac output, or other value related to arterial tone. (Denoted by 50). The same method can be used for other cardiovascular parameters.

  Determining whether the patient in these methods is in a normal or abnormal hemodynamic state involves applying a multivariate statistical (ie, Boolean) model to the arterial pressure waveform. Such a multivariate statistical model includes a first arterial pressure waveform data set from a test or reference patient in a first group that has experienced an abnormal hemodynamic state and a second group that has experienced a normal hemodynamic state. And a second arterial pressure waveform data set from a reference or reference patient. The multivariate Boolean model includes a first prescribed value (1) relating to the arterial pressure waveform of the first arterial pressure waveform data set, a second prescribed value (2) relating to the arterial pressure waveform of the second arterial pressure waveform data set, and Provides an output value corresponding to. If the output value is greater than the threshold value between the first specified value and the second specified value, it is determined that the patient is in an abnormal hemodynamic state as a result. In other words, this multivariate Boolean model is set to provide different output values for each input data set. Specifically, the multivariate Boolean model has a first specified value related to the arterial pressure waveform of the first arterial pressure waveform data set and a second specified value related to the arterial pressure waveform of the second arterial pressure waveform data set. Provide the corresponding output value.

  As a specific example, the determination of whether a patient's peripheral arterial pressure has been separated from the central aortic pressure can be determined by applying a multivariate statistical (ie, Boolean) model to the arterial pressure waveform data to Determining whether it has been separated from the aortic pressure. This multivariate Boolean model is based on the first arterial pressure waveform data set from the first group of patients undergoing separation of peripheral arterial pressure and central aortic pressure, and the separation of peripheral arterial pressure and central aortic pressure. And a second arterial pressure waveform data set from a second group of non-experienced test patients. The multivariate Boolean model is an output value corresponding to a first separation value relating to the arterial pressure waveform of the first arterial pressure waveform data set and a second separation value relating to the arterial pressure waveform of the second arterial pressure waveform data set. ("Separated output value"). If the separated output value is greater than the separation threshold between the first separation value and the second separation value, then it is determined that the patient's peripheral arterial pressure has been separated from the central aortic pressure.

  As used herein, a multivariate Boolean model is determined based on an element set that includes one or more parameters that are affected by an abnormal hemodynamic state. Without wishing to be fixed by theory, multiple elements are used to suggest a vascular condition that has occurred to the patient by utilizing a model and constraining the model to different output requirements for the first and second data sets. Can be used. Each type of factor used, such as the standard deviation of pulsations, typically records the difference between a patient experiencing a particular vascular condition and a patient not experiencing that condition. However, this difference often varies, and even if a patient is experiencing some vascular condition, a particular patient will have a value between a clear positive reading and a clear negative reading. For some reason, the patient's specific elements may be within the normal range. However, by using multiple elements, i.e., multiple elements that are affected by vascular conditions, an indication that is usually sufficiently positive (or not in that state) to indicate that it is in a particular state. A sufficiently negative reading to suggest). A multivariate Boolean model as described herein uses multiple elements to improve the ability to separately distinguish between two states, ie, those experiencing or not experiencing peripheral separation. There is ability.

  The particular number of elements used in the multivariate Boolean model is determined according to the ability of the individual elements to distinguish between patients in a particular state and patients not in a particular state. The number of elements may be further increased to provide the model with a high level of accuracy. Therefore, a large number of factors can be used to facilitate the model's precision, accuracy, and / or reproducibility, as required in certain situations. Examples of elements that can be used in the model described herein include parameters based on the standard deviation of pulsations in the arterial pressure waveform data set (a), parameters based on the R versus R interval in the arterial pressure waveform data set (heart rate (B), a parameter (c) based on the region under the cardiac contraction in the arterial pressure waveform data set, a parameter (d) based on the cardiac contraction period in the arterial pressure waveform data set, and a cardiac dilation in the arterial pressure waveform data set Parameter (e) based on period, parameter (f) based on arterial pressure average in arterial pressure waveform data set, parameter (g) based on standard deviation of pressure weight in arterial pressure waveform data set, pressure weight in arterial pressure waveform data set Parameter (h) based on the mean value of arterial beats in the arterial pressure waveform data set Parameter (i) based on the degree of skewness, parameter (j) based on the kurtosis of arterial pulsation in the arterial pressure waveform data set, parameter (k) based on the pressure-weighted skewness in the arterial pressure waveform data set, arterial pressure waveform A parameter (l) based on pressure-weighted kurtosis in the data set and a parameter (m) based on pressure-dependent wind-kessel compliance in the arterial pressure waveform data set are included. Additional elements that can be used with the multivariate statistical models described herein include the form of heart rate arterial pressure signals (such as time domain, frequency domain, or time frequency domain measurements) and one or more ranks. A parameter (n) based on at least one statistical moment of the arterial pressure signal having, a parameter corresponding to the heartbeat (o), and a parameter set (p) of patient anthropometry. One or more of these elements (or all of these elements) can be used in the multivariate statistical model described herein.

  The elements used in the multivariate Boolean and multivariate statistical models described herein are calculated from signals based on, or derived from, or a function of arterial pressure. Calculation of cardiovascular parameters such as arterial compliance (arterial tone) is disclosed in US patent application Ser. No. 10 / 890,887 (filed Jul. 14, 2004), the entire contents of which are hereby incorporated by reference. Embedded in. Examples of elements and data used to calculate cardiovascular parameters used in the methods disclosed herein, including the parameters disclosed in US patent application Ser. No. 10 / 890,887 are described below.

FIG. 5 shows an example of the arterial pressure waveform P (t) acquired over one cycle of the heartbeat. Here, from the point of the diastolic pressure P _dia at the time t _dia0 , the time t _sys of the systolic pressure P _sys is _calculated. Thus, the cycle until time t _dia1 when the blood pressure reaches P _dia again.

  Signals useful for use with the present method are proportional to arterial pressure, or arterial pressure measured invasively or non-invasively at any location in the arterial tree, such as the ribs, thighs, or upper arm. Cardiovascular parameters based on any signal that is, is derived from, or a function of. If an invasive means, in particular a pressure transducer with a catheter attached, is used, the result is that any position of the artery can be measured. Non-invasive transducer placement is typically determined by its own means such as finger cuffs, upper arm pressure cuffs, and earlobe clamps. Regardless of the particular means used, the resulting data ultimately produces an electrical signal that corresponds to (eg, is proportional to) arterial pressure.

As shown in FIG. 6, any standard analog-to-digital converter (ADC) can be used to digitize an analog signal, such as arterial pressure, into a sequence of digital values. In other words, using known methods and electrical circuits, the arterial pressure t ₀ ≦ t ≦ t _f can be converted to the digital form P (k), k = 0, (n−1). In this digital form, t ₀ and t _f are the first and last time of the measurement interval, n is the number of arterial pressure samples included in the calculation, and is usually assigned uniformly over the calculation interval.

In order to obtain relevant data from such a digital or digitized signal, an ordered collection method of m values, where i = 1 to (m−1) sequence Y (i), is considered. As is known in the field of statistics, the first four moments μ ₁ , μ ₂ , μ ₃ , and μ ₄ of Y (i) can be calculated using known equations. This is because μ ₁ is the mean value (ie arithmetic mean), μ ₂ = σ ² is the variation value (ie square value of the standard deviation σ), μ ₃ is skewness, and μ ₄ is kurtosis. Therefore, the following equation holds.
(Equation 1)
μ1 = Y _avg = 1 / m * Σ (Y (i)) (Formula 1)
(Equation 2)
μ ₂ = σ ² = 1 / (m−1) * Σ (Y (i) −Y _avg ) ² (Formula 2)
(Equation 3)
μ ₃ = 1 / (m−1) * Σ [(Y (i) −Y _avg ) / σ] ³ (Formula 3)
(Equation 4)
μ ₄ = σ / (m−1) * Σ [(Y (i) −Y _avg ) / σ] ⁴ (Formula 4)
In general, the β-th moment μ _p can be expressed as the following equation where i = 0 to (m−1).
(Equation 5)
μ _β = 1 (m−1) * 1 / σ ^β * Σ [(Y) (i) −Y _avg )] ^β (Formula 5)
The discrete value equations for the second through fourth moments are usually approximated using 1 / (m−1) instead of 1 / m for well-known statistical reasons.

  The method described herein can utilize elements that are functions of a pressure-weighted time vector as well as the four moments of the pressure waveform P (k). The standard deviation σ provides one level of shape information. That is, as σ becomes larger, the function Y (i) “spreads”, that is, the tendency to deviate from the average value increases. Standard deviation provides some shape information, but its drawbacks are easily understood by considering the following. That is, the order in which the values generating the sequence Y (i) are “reversed”, ie Y (i) is reflected and moved around the i = 0 axis, and the value Y (m−1) is in time In the case of the first value, the average value and the standard deviation do not change.

  The skewness is a measurement value that is incongruent and indicates that the left or right side of the function Y (i) related to the statistical mode is larger than the other. As the positively biased function rapidly increases and reaches its peak value, it then decreases slowly. This is true even for negatively biased functions. It includes shape information where the skewness is not found in the mean or standard deviation value, especially it suggests how quickly the function first reaches its peak value and then slowly decreases is important. Two different functions can have the same mean and standard deviation, but they rarely have the same skewness.

  Kurtosis is a measure of whether the function Y (i) is sharper or more uniform than the normal distribution. Thus, high kurtosis shows a clear peak value near the mean value, followed by a decrease, followed by a long “back”. Low kurtosis tends to suggest that the function is relatively uniform in the region of its peak value. The normal distribution has a kurtosis of 3.0, and the actual kurtosis is therefore often adjusted to 3.0, which is relative to the initial value.

  The advantage of using four statistical moments in the heart rate interval arterial pressure waveform is that the moment is accurate and provides a highly accurate mathematical measurement in the shape of the heart rate interval arterial pressure waveform. In cases where arterial compliance and peripheral resistance directly affect the shape of the arterial pressure waveform, the effects of arterial compliance and peripheral resistance may be directly evaluated by measuring the shape of the arterial pressure waveform at the heart rate interval. The statistical moment of heart rate waveform shape sensitivity in addition to the other arterial pressure parameters described herein is effective in measuring the combined effects of vascular compliance and peripheral resistance, i.e. arterial tone. May be used. Arterial tone exhibits a combined effect of arterial compliance and peripheral resistance, and corresponds to the impedance of the model comparable to the two-component electrical analog of the known Wind Kessel hemodynamic model consisting of capacitive and resistive components. By measuring arterial tone, several other parameters based on arterial tone such as arterial elasticity, stroke volume, and cardiac output can also be measured directly. Any of those parameters may be used as an element in the methods described herein.

In calculating the first four moments μ _1p , μ _2p , μ _3p , and μ _4p of the pressure waveform P (k), a multivariate Boolean or multivariate statistical model is used. Here, μ _1p is the mean value, μ _2p P = σ _p ² is the variation value, that is, the square value of the standard deviation σ _p , μ _3p is the skewness, and μ _4p is the kurtosis, and all these moments are It is determined based on the pressure waveform P (k). Equations 1-4 above can be used to calculate these values, substituting P for Y instead of k, k for i, and n for m.

Equation 2 above provides an “exemplary” method for calculating the standard deviation. Other more approximate methods can also be used. For example, at least in the context of blood pressure based measurements, the approximate σ _p is determined by dividing the difference between the measured maximum pressure value and the minimum pressure value by 3, and the maximum of the first derivative of P (t) with respect to time The absolute value of the value, or minimum, is usually proportional to σ _p .

As shown in FIG. 6, the measured pressure corresponding to each individual time k is P (k). The values k and P (k) can be converted into a sequence T (j) corresponding to a column diagram, meaning that each P (k) value is used as a “number” of the corresponding k values. According to a very simplified example, the overall pressure waveform consists of only four measured values P (1) = 25, P (2) = 50, P (3) = 55, and P (4) = 35 Is guessed. This is then represented as a sequence T (j) which is 25 1, 50 2, 55 3, and 35 4. This means
T (j) = 1, 1, 2, 1, 2, 2, 3, 3, 3, 4, 4, 4
And this sequence therefore has a duration of 25 + 50 + 55 + 35 = 165.

  Moments are calculated in this sequence like any other. For example, the average value (first moment) is calculated by the following equation.

(Equation 6)
μ _1T = (1 * 25 + 2 * 50 + 3 * 55 + 4 * 35) /165=430/165=2.606 (formula 6)
Here, the standard deviation σ _T is the square root of the fluctuation value μ _2T , and the following equation is established.
SQRT [1/164 * 25 (1-2.61) ² +50 (2-2.61) ² +55 (3−2.61) ² +35 (4-2.61) ² ] = 0.985
The skewness μ _3T and kurtosis μ _4T can be calculated by similar substitute values in equations 3 and 4.
(Equation 7)
μ _3T = {1 / (164) * (1 / σ _T ³ ) Σ [P (k) * (k−μ _1T ) ³ ]} (Expression 7)
(Equation 8)
μ _4T = {1 / (164) * (1 / σ _T ⁴ ) Σ [P (k) * (k−μ _1T ) ⁴ ]} (Formula 8)
Here, k = 1 to (m−1).

As shown in these equations, this process actually “weights” the value k at each individual time by its corresponding pressure value P (k) before calculating the moment of time. The sequence T (j) has very useful properties that firmly characterize the timing distribution of the pressure waveform. Reversing the order of the pressure values P (k) changes in most cases not only all higher-order moments but also the average value of T (j). In addition, secondary “humps” that typically occur in heavy beat pressure _Pdicotic also have a significant impact on kurtosis μ _4T . On the other hand, simply identifying overlapping ridges in prior art such as the Romano method requires a complex calculation of at least one derivative.

  The pressure-weighted moment provides another level of shape information in the heart-rate arterial pressure signal, where both the amplitude and time information of the heart-rate arterial pressure signal are very accurately measured. By using pressure-weighted moments in addition to pressure waveform moments, the accuracy of the model described herein can be improved.

  One useful cardiovascular parameter for use with the methods described herein may be used as a cardiovascular parameter per se or other cardiovascular such as stroke volume or cardiac output. There is an arterial tone factor χ that can be used to calculate the parameters. All four values of pressure waveform and pressure weighted time moment are used to calculate arterial tone χ. Additional values are included in calculations that take into account other known characteristics such as patient-specific complex types in vascular bifurcation. Examples of further values include heart rate HR (or duration of R wave), body surface area BSA or other anthropometric parameters in the patient, such as Langewaters ("The Static Elastics of 45 Human Thoracic and 20 Abnormal Aortas in Vitro"). the Parametrs of a New Model "J. Biomechanics, 17 (6): 425-435 (1984)), etc., a known method for calculating compliance as a polynomial function of pressure waveform and patient age and gender. At least one statistical moment of the arterial pressure signal having a compliance value C (P), calculated using Parameter based on, it includes parameters based on the region of the heart contractions subordinates arterial pressure signal, based on the ratio of the parameters based on the systolic period, and systolic period and diastolic period parameter.

  The last three cardiovascular parameters mentioned above, namely the area under the systole of the arterial pressure signal, the systolic period, and the ratio between the systolic period and the diastole period are influenced by arterial tension and vascular compliance, and thus for example Different between patients in normal hemodynamic status and patients in high cardiac output status. Because these three cardiovascular parameters are different between normal and high cardiac output patients, the methods described herein use these cardiovascular parameters to vascularize the patient's peripheral arteries. Dilation or vasoconstriction can be detected.

The graph of FIG. 7 shows the region under the heart contraction part of the arterial pressure waveform (A _sys ). The region under the cardiac contraction in the arterial pressure waveform in the arterial pressure signal is defined as the region under the waveform that starts from the beginning of the pulse and ends at the stage of the heavy staying pulse (from point b to point d in FIG. 7). The The area under cardiac contraction represents the energy of the arterial pressure signal during cardiac contraction, which is directly proportional to stroke volume and inversely proportional to arterial compliance. In measuring the entire group of normal patients and patients with high cardiac output, the movement of A _sys can be detected. As shown in FIG. 8, the energy of the arterial pressure signal during cardiac contraction is higher in some patients, for example, in a high cardiac output state. Those patients with higher A _sys are usually patients with increased CO, ie, high cardiac output (CO), and low or normal HR, mainly due to increased cardiac contractility. This means that those patients have increased stroke volume and reduced arterial compliance that are directly reflected in the energy of the arterial pressure signal during cardiac contraction. This reflected wave, which is usually very strong during many high cardiac output conditions, can significantly affect the increased energy in the signal during cardiac contraction.

The cardiac contraction period (t _sys ) is shown in the graph of FIG. The cardiac contraction period in the arterial pressure waveform is defined as the period from the start of the pulse to the stage of the heavy dormant pulse (from point b to point d in FIG. 9). The systolic period is directly affected by arterial compliance, but is relatively unaffected by changes in peripheral arterial tone, except when large reflected waves are present. As shown in FIG. 10, for example, the cardiac contraction period of some high cardiac output patients is longer than the normal patient cardiac contraction period (data moving to a higher t _sys value). As shown as cardiac contraction energy, the systolic period is usually longer in patients with high CO who also have low or normal HR. Here, CO is mainly enhanced by increased cardiac contractility, and the contractility cannot be high enough to increase cardiac contractile energy. The increased stroke volume of those patients is due in part to increased contractility and cardiac contraction duration. The reflected wave has the same effect.

Additional parameters that differ between normal and high cardiac output patients include, for example, the ratio between systolic period (t _sys ) and diastole period (t _dia ), as shown in the graph of FIG. There is. The period of diastole in the arterial pressure waveform is defined as the period from the stage of the dormant pulse to the end of the heart cycle (from point d to point e in FIG. 11). In some high cardiac output conditions, the ratio between the period of systole and diastole is significantly greater than that observed in normal hemodynamic conditions. This is usually observed in septic shock patients with elevated CO, also high HR. In these types of conditions, the cardiac contraction occupies almost the entire heart cycle except for a very short period for diastole before the next heart cycle begins. This is illustrated in FIG. 12, which shows the diastole period during high HR conditions in septic shock patients and normal patients, and in FIG. 13, which shows the systole period. As shown in these figures, high HR patients in normal hemodynamic state (dashed lines) tend to have short periods of both cardiac contraction and diastole while high in septic shock HR patients (thick line) tend to have a short diastole period, although with normal or long systole periods.

  Other parameters based on arterial tone factors such as stroke volume (SV), cardiac output (CO), arterial flow, or arterial elasticity may be used as factors in the methods described herein. In one example, stroke volume (SV) may be calculated as the product of arterial tone and the standard deviation of the arterial pressure signal.

ここで、ＳＶは一回拍出量であり、χは動脈緊張であり、そしてσ_ｐは動脈圧の標準偏である。

Where SV is stroke volume, χ is arterial tone, and σ _p is the standard deviation of arterial pressure.

Over the time period during which each calculation is performed, the analog measurement interval where the time window [t ₀ , t _f ] and therefore the individual sample interval k = 0 to (n−1) is preferably small enough, Does not cause substantial movement in pressure and / or time moment. However, a time window that extends longer than one heart cycle provides adequate data. The measurement interval is preferably a plurality of heartbeat cycles where different heartbeat cycles start and end at the same point. By using a plurality of heartbeat cycles, it is possible to reliably use a pressure average value P _{avg in} which the pressure average value used for calculation of various higher-order moments is not biased due to an incomplete measurement of one cycle.

  Larger sample windows have the advantage that perturbation effects caused by reflection, for example, are usually reduced. The appropriate time window can be determined using standard experimental and clinical methods known to those skilled in the art. Note that the time window can coincide with a single heartbeat cycle that is not affected by the average pressure movement.

The time window [t ₀ , t _f ] can also be adjusted according to the trend of P _avg . For example, if P _avg over a given time window differs completely or proportionally from the previous time window P _avg by more than a threshold, the time window may be reduced as a result. In this case, it can be suggested that the stability of P _avg can result in an expanded time window. The time window can also be scaled up or down based on the noise source or signal-to-noise ratio or varying measurements. Preferably, a limit is set regarding the extent to which the time window is allowed to expand or contract, and if all such expansion or contraction is allowed, then the time interval is preferably displayed to the user.

The time window need not start from any particular position in the heart cycle. Thus, t ₀ need not be the same as t _dia0 , although in many implementations this can be a convenient choice. Thus, the start and end of each measurement interval (ie, t ₀ and t _f ) can be triggered by almost all characteristics of the heart cycle, eg, time t _dia0 or t _sys , or R wave _uncompressed characteristics, for example.

  Rather than directly measuring blood pressure, any other input signal that is proportional to, derived from, or a function of blood pressure can be used. This means that calibration can be performed at any or all of the multiple stages in the calculation. For example, if a signal other than arterial pressure itself is used as input, it is calibrated to blood pressure before its value is used to calculate various component moments, or is calibrated afterwards, in which case Each moment value produced can be increased or decreased. That is, the fact that cardiovascular parameters can use different input signals in some cases than directly measuring arterial pressure does not impede the ability to generate accurate compliance values.

  In one example, the isolated output values for the multivariate Boolean model in a given set of elements are constrained to a first separation value for those elements derived from the arterial pressure waveform of the first arterial pressure waveform data set, and simultaneously , Constrained to a second separation value for those elements obtained from the arterial pressure waveform of the second arterial pressure waveform data set. The first separation value and the second separation value may be set to provide a convenient comparison for the patient being analyzed to the separated output value. For example, the first separation value may be greater than the second separation value. Further, the first separation value may be a positive number and the second separation value may be a negative number. In a further example, the first separation value may be +100 and the second separation value may be −100.

Such separation values, ie +100 and -100, can be used to generate Boolean results in a multivariate Boolean model. For example, in a multivariate Boolean model used with a first separation value of +100 and a second separation value of -100, the separated output value can be set using the following indicators:
Separated output ≧ 0 → vascular condition is shown.
Separated output <0 → vascular condition is not shown.
In this case, a “0” value is considered as a threshold above which separation is indicated, and a value greater than zero was used to build a multivariate model rather than a patient who did not experience separation. It can be shown to be closely related to the patient who experienced the separation. Conversely, a value less than “0” indicates a patient experiencing segregation more closely related to a patient who did not experience the segregation used to build the multivariate model than to a patient who experienced segregation. . In this example, the threshold is set to the average value between the first separation value and the second separation value. However, the threshold can be changed based on empirical observations. For example, if the value α is an average value between the first threshold value and the second threshold value, the resulting threshold values are α-1, α-2, α-3, α-4, α-5, It may be α-10, α-15, α-20, α + 1, α + 2, α + 3, α + 4, α + 5, α + 10, α + 15, or α + 20.

In a multivariate Boolean model as described herein, a threshold range that is not critical to whether a patient is experiencing separation or other vascular conditions may be used. In such a case, a threshold range may be used. For example, in a multivariate Boolean model used with a first separation value of +100 and a second separation value of −100, the model output may be set using the following indicators:
Separated output ≧ 10 → vascular condition is shown.
−10 <separated output <10 → indeterminate (continue analysis)
Separated output ≦ −10 → vessel condition is not shown.
In this example, values between -10 and 10 are considered as indeterminate, and further data may be analyzed to determine whether a value greater than 10 or a value less than -10 is indicated. Alternatively, in this situation, an isolated output value greater than or equal to 10 indicates isolation, and a value less than or equal to −10 indicates a normal vascular condition. The threshold range is determined based on the evaluation of the model capability so as to indicate the vascular condition at an intermediate point between the first separation value and the second separation value. In addition, the threshold range can be changed and is otherwise adjusted based on empirical observations. For example, when the value α is an average value between the first threshold value and the second threshold value, the resulting threshold values are [α-1 ± β], [α-2 ± β], [α-3]. ± β], [α-4 ± β], [α-5 ± β], [α-10 ± β], [α-15 ± β], [α-20 ± β], [α + 1 ± β], [Α + 2 ± β], [α + 3 ± β], [α + 4 ± β], [α + 5 ± β], [α + 10 ± β], [α + 15 ± β], [α + 20 ± β] may be used. Here, β is the upper and lower limits of the range, for example β may be 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 15, or 20 depending on the model.

  The generation of a multivariate Boolean model involves several steps. For example, the response surface of multiple linear regression can be used to build a model. The number of conditions used in the model can be determined using several numerical approaches to minimize the mean square error between the model output value and the value that the model is forced for a particular condition. As a more detailed example, the predictor set in the model output value is associated with a particular group of patients. That is, the default value may be set to +100 for patients experiencing a particular vascular condition and -100 for patients not experiencing a vascular condition. This operation generates a series of reference values, each of which is a function of the component parameter of the model output value. The multivariate approximation function can then be calculated using known numerical methods that best relate the model output to +100 or -100 depending on the group of patients in a defined manner. The polynomial multivariate adjustment function may then be used to generate polynomial coefficients that meet the +100 and −100 constraints in each set of predicted values. Such a multivariate model (β is the model output) has the following general form:

ここで、α_１〜α_ｎは多項式の重回帰モデルの係数であり、ｘ_１〜ｘ_ｎはモデルの予測変数である。予測変数は動脈圧波形から算出された前述の要素から選択される。

Here, α _{1 to} α _n are coefficients of a polynomial multiple regression model, and x _{1 to} x _n are predictive variables of the model. The predictive variable is selected from the aforementioned elements calculated from the arterial pressure waveform.

Each predictor variable x _{i of the} model is predefined from a combination of arterial pressure waveform parameters v _i and can be calculated as follows:

係数ａ_ｉおよび指数行列「Ｐ」は、患者から収集された要素データを用いて多変量最小二乗回帰により決定され得る。それらは例えば、基準患者の個体数における患者の群に応じた＋１００および−１００の値に関連し得る。例えば、分離を経験している患者からの要素データは、次式に関連し得る。

The coefficients a _i and the exponent matrix “P” can be determined by multivariate least squares regression using elemental data collected from the patient. They can be associated with values of +100 and -100 depending on the group of patients in the reference patient population, for example. For example, element data from a patient experiencing separation may be related to the following equation:

分離を経験していない患者からの要素データは、次式に関連し得る。

Element data from patients not experiencing separation can be related to the following equation:

多変量ブールモデルの具体例では、モデルのための最良適合を提供するのに１７項を要求する１４の動脈圧波形要素（Ｘ_ｉ）を用いて生成された。これらのパラメータには、ｖ_１（拍動の標準偏差（ｓｔｄ））、ｖ_２（Ｒ対Ｒ間隔（ｒ２ｒ））、ｖ_３（心臓収縮下の領域（ｓｙｓ 領域））、ｖ_４（心臓収縮期間（ｔ＿ｓｙｓ））、ｖ_５（心臓収縮全体の期間（ｔ＿ｄｉａ））、ｖ_６（動脈圧平均値（ＭＡＰ））、ｖ_７（圧力加重の標準偏差（σ_Ｔ））、ｖ_８（圧力加重の平均値（μ_２））、ｖ_９（動脈拍動の歪度（μ_３Ｐ））、ｖ_１０（動脈拍動の尖度（μ_４Ｐ））、ｖ_１１（圧力加重の歪度（μ_３Ｔ））、ｖ_１２（圧力加重の尖度（μ_４Ｔ））、ｖ_１３（圧力依存性のウインドケッセルコンプライアンス（Ｃ_Ｗ））、およびｖ_１４（患者の体表面積（ＢＳＡ））がある。モデルは下記の式である。

In the multivariate Boolean model embodiment, it was generated with 14 arterial pressure waveform elements (X _i ) requiring 17 terms to provide the best fit for the model. These parameters include: v ₁ (standard deviation of pulsation (std)), v ₂ (R vs. R interval (r2r)), v ₃ (region under systolic (sys region)), v ₄ (cardiac contraction) Period (t_sys)), v ₅ (total systolic period (t_dia)), v ₆ (arterial pressure mean value (MAP)), v ₇ (standard deviation of pressure weight (σ _T )), v ₈ (pressure weighted) Mean value (μ ₂ )), v ₉ (arterial pulsation skewness (μ _3P )), v ₁₀ (arterial pulsatility (μ _4P )), v ₁₁ (pressure-weighted skewness (μ _3T) )), V ₁₂ (pressure-weighted kurtosis (μ _4T )), v ₁₃ (pressure-dependent wind-kessel compliance (C _W )), and v ₁₄ (patient body surface area (BSA)). The model is:

回帰後、配列Ｐ（１７×１４）の値が決定されて、モデルに含まれる変数を下記のように規定する。

After the regression, the value of the array P (17 × 14) is determined, and the variables included in the model are defined as follows.

２以下の次数を有する各パラメータを用いて、回帰条件ごとのパラメータの数が３以下に制限されるように回帰が実行された。したがって前述のように、行列Ｐの各行は最大２であるＰの各要素の絶対値と共に、最大３のノンゼロ項を有する。これらの制約は数値安定性および正確性を目的として設定された。β用の式は、それ故、１４次元のパラメータ空間における１７項２次曲線となった。β用に生成された式を下記に示す。

Regression was performed using each parameter having an order of 2 or less so that the number of parameters per regression condition was limited to 3 or less. Thus, as described above, each row of matrix P has a maximum of 3 non-zero terms, with an absolute value of each element of P being a maximum of 2. These constraints were set for numerical stability and accuracy. The equation for β was therefore a 17-term quadratic curve in a 14-dimensional parameter space. The formula generated for β is shown below.

ここで、ａ_１〜ａ_１７は多項式の複数回帰モデルの係数である。これらの係数は、次に、選ばれた一連の要素を与えられたβが＋１００に等しくなるようにパラメータを最もうまく関連させるように（公知の数値方法を用いて）計算された。この場合では、選ばれた要素は波形パラメータｖ_ｊ（前述で規定した）に基づく式１４により決定された。最小二乗回帰に基づく多項式の多変量調整関数が、多項式用の下記の係数を生成するのに用いられた。

Here, a _{1 to} a ₁₇ are coefficients of a polynomial multiple regression model. These coefficients were then calculated (using known numerical methods) to best relate the parameters such that given a set of elements given β equals +100. In this case, the selected element was determined by Equation 14 based on the waveform parameter v _j (defined above). A polynomial multivariate adjustment function based on least squares regression was used to generate the following coefficients for the polynomial:

このような多変量ブールモデルが改良されると、分離された状態のいかなる患者もリアルタイムで連続的に検出するのに用いられ得る。このモデルは患者の動脈圧波形から決定された選ばれた要素を用いて連続的に評価され得る。この例では、第１の分離値は＋１００に設定されて、第２の分離値は−１００に設定された。これにより閾値は、例えば、次式を提供するようにこのモデルにおいてゼロに設定された。

When such a multivariate Boolean model is improved, any patient in isolation can be used to continuously detect in real time. This model can be continuously evaluated using selected elements determined from the patient's arterial pressure waveform. In this example, the first separation value was set to +100 and the second separation value was set to -100. This caused the threshold to be set to zero in this model, for example, to provide:

  β ≧ 0 → indicates separation.

  β <0 → indicates no separation.

  In the method of the present application, if it is determined that the patient has been separated, it is then determined whether the patient has a higher cardiac output. Determining whether the patient has high cardiac output includes applying a multivariate Boolean model of high cardiac output to the arterial pressure waveform data. This high cardiac output multivariate boule model uses the same technique as described above for the isolated multivariate boule model, using a first arterial pressure from a first group of patients with high cardiac output. Created from a waveform data set and a second arterial pressure waveform data set from a second group of patients who were not at high cardiac output. The multi-variable Boolean model of high cardiac output is a first high cardiac output value of the arterial pressure waveform in the first arterial pressure waveform data set and a first of the arterial pressure waveform in the second arterial pressure waveform data set. Providing an output value of high cardiac output corresponding to two high cardiac output values. If the output value of the high cardiac output is equal to or higher than the threshold value of the high cardiac output between the specified value of the first high cardiac output and the specified value of the second high cardiac output, Results The patient is determined to have high cardiac output. The first and second high cardiac output values and the high cardiac output threshold are determined and / or defined in the same manner as defined for the multivariate Boolean model in which similar values are separated.

  In one example of the method of the present application, if patient isolation and high cardiac output are determined, a group of test patients who have experienced high cardiac output status and isolation between central and peripheral arterial pressure High cardiac output multivariate statistical model generated using arterial pressure waveform data from to determine a patient's high cardiac output and isolated arterial tone (or other cardiovascular parameters) Used. Similarly, the present method experiences high cardiac output status and separation between central and peripheral arterial pressure if the patient is not determined to be isolated or is not determined to have high cardiac output. A normal multivariate statistical model generated using arterial pressure waveform data from a group of unexamined test patients is used to determine the patient's normal arterial tone (or other cardiovascular parameters). A normal high cardiac output multivariate statistical model is generated using the same method as described below.

Generation of a multivariate statistical model for calculating arterial tone, which is an example of a cardiac parameter, is accomplished in the same manner as generating a multivariate Boolean model as described above. The main difference is that the model χ is not constrained to any specified value. Rather, the multivariate model χ in arterial tone is, in some pre-defined sense, a plurality of parameters in the model χ arterial pressure waveform, which is based on a series of SV and arterial pressure measurements. It is calculated using a known numerical method that best relates to tension (as defined in Equation 9 as SV / σ _p ). Specifically, a polynomial multivariate adjustment function is used to generate a coefficient of a polynomial that gives a χ value in each arterial pressure waveform parameter set as described below.

ここで、ａ_１〜ａ_ｎは多項式の重回帰モデルの係数であり、χ_１〜χ_ｎはモデルの予測変数である。予測変数は、動脈圧波形から算出された前述の要素から選択される。

_Here, a 1 ~a _n are the coefficients of the multiple regression model polynomial, χ ₁ ~χ _n is predictor model. The predictive variable is selected from the aforementioned elements calculated from the arterial pressure waveform.

Each of the model predictor variables χ _i is a predefined combination in the arterial pressure waveform parameter v _i and can be calculated by the following equation.

係数ｖ_ｉは動脈圧波形の異なる時間および周波数領域パラメータである。

The coefficient v _i is a different time and frequency domain parameter of the arterial pressure waveform.

An approximation of the model is achieved using reference measurements in a group of reference patients. The arterial tone model χ _decoupled used for patients with high cardiac output experiencing peripheral arterial pressure separation is a normal patient and patients with high cardiac output experiencing peripheral pressure separation. Constructed with a reference data set containing measurements from both. The arterial tone model χ _normal used for both patients in normal hemodynamic state is constructed using a reference data set containing measurements from patients in normal state.

A specific example of a multivariate statistical model was generated using 11 arterial pressure waveform parameters. These parameters are v ₁ (standard deviation of pulsation (σ _P )), v ₂ (heart rate), v ₃ (mean value of arterial pressure (P _avg )), v ₄ (standard deviation of pressure weight (σ _T) )), V ₅ (pressure weighted MAP (μ _1T )), v ₆ (arterial pulsation skewness (μ _3P )), v ₇ (arterial pulsatility (μ _4P )), v ₈ (pressure weighted) Skewness (μ _3T )), v ₉ (pressure-weighted kurtosis (μ _4T )), v ₁₀ (pressure-dependent wind-kessel compliance (C _W )), and v ₁₁ (patient body surface area (BSA)) The coefficient a _i and the exponential matrix “P” can be determined by multivariate least squares regression using elemental data collected from the patient, and the coefficient and exponential element can be determined according to the number of individuals in the reference patient. Related to the “exact” stroke volume determined using the dilution method, this model A And P were determined by the following equations.

回帰は２以下の次数を有する各パラメータを用いて、３未満の回帰変数につきパラメータの数を制限する方法により実行される。したがって行列Ｐの各行は、最大２である各要素Ｐの絶対値と共に、最大３つのノンゼロ項を有する。これらの制約が数値安定性および正確性を達成するために用いられた。χ用の式はそれ故、１１次元のパラメータ空間における２次曲線となった。χ用に決定された多項式の式は下記に記載される。

Regression is performed by a method that limits the number of parameters per regression variable of less than 3 using each parameter having an order of 2 or less. Thus, each row of matrix P has a maximum of three non-zero terms, with an absolute value of each element P being a maximum of two. These constraints were used to achieve numerical stability and accuracy. The formula for χ was therefore a quadratic curve in an 11-dimensional parameter space. The polynomial equation determined for χ is described below.

したがって、患者の心臓血管パラメータはこのような（すなわち、（モデルに応じて）心臓血管パラメータに応じた血圧パラメータを表す臨床的に生成された基準測定値セットに関連する近似関数、前述の１つ以上のパラメータの関数である近似関数、および正常な血行動態状態にある患者または異常な血行動態状態にある患者からの心臓血管パラメータに応じた血圧パラメータを表す臨床的に決定された基準測定値セットを決定する）第１の特定のモデルにより決定され得る。次に、動脈圧波形データから動脈圧パラメータセットを決定する。この動脈圧パラメータセットは、多変量統計モデルの生成に用いられるのと同じパラメータを含む。次に、動脈圧パラメータセットを用いて近似関数を評価することにより患者の心臓血管パラメータを計算する。

Thus, the patient's cardiovascular parameter is such an approximation function associated with a clinically generated reference measurement set representing a blood pressure parameter as a function of the cardiovascular parameter (depending on the model), one of the aforementioned Approximate function that is a function of the above parameters, and a clinically determined set of reference measurements that represent blood pressure parameters as a function of cardiovascular parameters from patients in normal or abnormal hemodynamic conditions Can be determined by a first specific model. Next, an arterial pressure parameter set is determined from the arterial pressure waveform data. This arterial pressure parameter set includes the same parameters used to generate the multivariate statistical model. The patient's cardiovascular parameters are then calculated by evaluating the approximation function using the arterial pressure parameter set.

  These methods for determining a patient's cardiovascular parameters can be used for continuous calculation of the patient's cardiovascular parameters, thereby accounting for changes in conditions over time. This method further alerts the user that the abnormal condition has been determined. This alert may be a notification or sound in a graphical user interface.

  The major components of a system that implements the methods described herein for determining cardiovascular parameters, such as patient arterial tone, are shown in FIG. This method may be implemented with an existing patient monitoring device, which may be a dedicated monitor. As described above, pressure, or some other input signal that is proportional to, derived from, or a function of pressure is actually in one or both of two ways: invasive and non-invasive Can be detected. For convenience, the system is described as measuring arterial pressure rather than as some other input signal that is converted to pressure.

  Both types of pressure sensing devices are shown in FIG. 14 for completeness purposes. In most practical applications of the methods described herein, one or several modifications are usually performed. In invasive applications in the methods described herein, a conventional pressure sensor 100 is attached to a catheter 110 that is inserted into an artery 120 of a body part 130 of a patient or diseased animal. The artery 120 is any artery in the arterial system, such as an artery in the thigh, rib or upper arm. For non-invasive applications in the methods described herein, conventional methods such as photoelectric pulse wave blood pressure probes may be used by any conventional method, such as using a sphygmomanometer band around the finger 230 or a transducer attached to the patient's wrist. The pressure sensor 200 is attached from the outside. Both types are shown schematically in FIG.

  Signals from sensors 100, 200 are sent through any known connector as an input to processing system 300. The processing system 300 typically includes process signals and includes one or more processors that execute code and other supporting hardware and system software (not shown). The methods described herein may be implemented using modifications, standards, personal computers, or may be incorporated into larger specialized monitoring systems. The processing system 300 for use with the methods described herein may further be connected to or include a conditioning circuit 302 that performs normal signal processing tasks such as amplification, sorting, or ranging as required. Good. The regulated, detected input pressure signal P (t) is then converted to digital form by a conventional analog to digital converter ADC 304 having a time reference or obtained from the clock circuit 305. As is known, the sampling frequency of the ADC 304 can be chosen in relation to the Nyquist criterion, thereby avoiding pressure signal aliasing (this procedure is very well known in the field of digital signal processing). ). The output from the ADC 304 is an individual pressure signal P (k) whose value can be stored in a conventional storage circuit (not shown).

  The value P (k) is accessed from or sent to memory using software module 310. This module 310 comprises computer-executable code for implementing a multivariate Boolean model that determines whether a patient has peripheral isolation and / or high cardiac output. The design of such a software module 310 is easy for those skilled in computer programming.

  For example, if patient specific data such as age, height, weight, BSA is used, they are stored in a memory area 315 that also stores other predetermined parameters such as threshold values or threshold range values. These values can be entered by conventional methods using any known input device 400.

  Calculation of arterial tone using an appropriate multivariate statistical model is performed in module 320. The calculation module 320 contains computer executable code, inputs the output of the module 310, and then performs the selected arterial tone calculation.

  As shown in FIG. 14, this result can be sent to the module (330) for further processing, and finally displayed on a conventional display or recording device 500 for presentation and understood by the user. The Similar to the input device 400, the display device 500 is similarly used by the processing system for other purposes.

  In each of the methods described herein, the user is notified when separation is detected. The user may be informed of the separation by notification on the display device 500 or other graphical user interface device. Sound may also be used to notify the user of separation. Both visual and auditory signals can be used.

  Exemplary embodiments of the present invention have been described with reference to block diagrams and flowchart illustrations of methods, apparatus, and computer program products. It will be appreciated by those skilled in the art that each block of the block diagram and flowchart description, and each combination of block diagram block and flowchart description, can be implemented by various means including computer program instructions. These computer program instructions are general purpose computers, special purpose computers, or instructions that cause a computer or other programmable data processing device to execute instructions that perform the functions specified in the blocks or blocks of the flowchart. It can be read into other possible data processing devices that produce the machine.

  The methods described herein are stored in a computer readable memory that can instruct a computer or other programmable data processing device such as a processor or processing system (shown at 300 in FIG. 14) to perform the functions in a particular method. Further obtaining computer program instructions to obtain. Thus, a product including computer readable instructions for executing the functions specified in the block shown in FIG. 14 is generated by the instructions stored in the computer readable memory. In a computer, processing system 300, or other programmable data processing apparatus to generate computer-executable processing in a series of operational steps performed on the computer, processing system 300, or other programmable device. In addition, computer program instructions can be read. This causes the instructions to be executed on a computer or other programmable device that provides the steps for performing the functions specified in the block. Also, the various software modules 310, 320, and 330 used to perform the various calculations and related method steps described herein can further be stored on a computer-readable medium as computer-executable instructions. Thereby, the method is executed and read by a different processing system.

  Thus, the block diagram block and flowchart descriptions establish a combination of means for performing a particular function, a combination of steps for performing a particular function, and a program instruction means for performing a particular function. Special purpose hardware-based computer system in which each block of the block diagram and the description of the flowchart and the combination of the block diagram and the description of the block diagram perform a specific function or process, or a combination of special purpose hardware and computer instructions It will be appreciated by those skilled in the art that it can be implemented in

  The present invention is not to be limited in scope by the embodiments disclosed herein that are intended as descriptions of some aspects of the invention and any embodiment that is functional equivalent within the scope of the invention. . Various modifications of the method shown and added to the described embodiments will be apparent to those skilled in the art and are intended to be within the scope of the appended claims. Further, only specific exemplary combinations of the method steps disclosed herein are specifically disclosed in the foregoing embodiments, but other combinations in the method steps will be apparent to those skilled in the art and are further appended. It is intended to be within the scope of the claims. Therefore, combinations of steps can be explicitly disclosed herein. However, other combinations of steps are included in the present invention even if not explicitly stated. As used herein, the term “comprising” and variations thereof are used synonymously with the term “including” and variations thereof, are open terms, and are not restricted terms.

Claims (48)

A method for calculating a patient's cardiovascular parameters comprising:
Providing the patient's arterial pressure waveform data;
Analyzing the arterial pressure waveform data to determine whether the patient is in an abnormal state;
If it is determined that the patient is in an abnormal state, applying a second multivariate statistical model to the arterial pressure waveform data to determine the patient's cardiovascular parameters;
Applying a third multivariate statistical model to the arterial pressure waveform data to determine the patient's cardiovascular parameters if the patient is not determined to be in an abnormal state.   Determining whether the patient is in an abnormal condition includes applying a first multivariate statistical model to the arterial pressure waveform data to determine whether the patient is in an abnormal condition, One multivariate statistical model includes a first arterial pressure waveform data set from a first group of test patients that were in an abnormal state and a second artery from a second group of test patients that were not in an abnormal state. And the first multivariate statistical model is a first value in the arterial pressure waveform of the first arterial pressure waveform data set and the second arterial pressure waveform data set. Providing an output value corresponding to a second value in the arterial pressure waveform, wherein the patient is in an abnormal condition if the output value is greater than a threshold value between the first value and the second value; The method of claim 1, wherein the method is determined to be. Determining whether the patient is in an abnormal state is
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on arterial pressure average value in the arterial pressure waveform data set, pressure in the arterial pressure waveform data set A parameter (g) based on a standard deviation of weights, a parameter (h) based on an average value of pressure weights in the arterial pressure waveform data set, a parameter (i) based on the skewness of arterial pulsations in the arterial pressure waveform data set, Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform data Parameter (k) based on pressure-weighted skewness in the set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set (M) and an approximate function that is a function of the parameter (n) based on the body surface area of the patient, a first arterial pressure waveform data set from a first group of test patients in an abnormal state; Determining an approximation function associated with a second arterial pressure waveform data set from a second group of test patients that were not in an abnormal state;
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on the arterial pressure average value in the arterial pressure waveform data set, Parameter (g) based on standard deviation of pressure weight, parameter (h) based on average value of pressure weight in arterial pressure waveform data set, parameter (i) based on arterial pulsation skewness in arterial pressure waveform data set Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform Parameter (k) based on pressure-weighted skewness in the data set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set Determining the arterial pressure parameter set from the arterial pressure waveform data comprising (m) and a parameter (n) based on the patient's body surface area;
Applying the first multivariate statistical model using evaluating the approximation function using the arterial pressure parameter set.   The method of claim 1, wherein the second multivariate statistical model is generated from an arterial pressure waveform data set from a group of test patients in an abnormal condition and provides values for the patient's cardiovascular parameters.   5. The method of claim 4, wherein the second multivariate statistical model is based on a set of elements that includes one or more parameters that are affected by the abnormal condition.   The method of claim 1, wherein the third multivariate statistical model is created from an arterial pressure waveform data set from a group of test patients who were not in an abnormal state and provides values for the patient's cardiovascular parameters.   The method of claim 6, wherein the third multivariate statistical model is based on a set of elements that includes one or more parameters used to calculate the cardiovascular parameters. The patient's cardiovascular parameters are:
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on arterial pressure average value in the arterial pressure waveform data set, pressure in the arterial pressure waveform data set A parameter (g) based on a standard deviation of weights, a parameter (h) based on an average value of pressure weights in the arterial pressure waveform data set, a parameter (i) based on the skewness of arterial pulsations in the arterial pressure waveform data set, Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform data Parameter (k) based on pressure-weighted skewness in the set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set (M) an approximation function that is a function of the parameter (n) based on the body surface area of the patient, the approximation relating to a clinically generated reference measurement set representing a blood pressure parameter according to the cardiovascular parameter Determining a function and a clinically determined reference measurement set representing blood pressure parameters in response to the cardiovascular parameters from a patient experiencing an abnormal condition;
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on the arterial pressure average value in the arterial pressure waveform data set, Parameter (g) based on standard deviation of pressure weight, parameter (h) based on average value of pressure weight in arterial pressure waveform data set, parameter (i) based on arterial pulsation skewness in arterial pressure waveform data set Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform Parameter (k) based on pressure-weighted skewness in the data set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set Determining an arterial pressure parameter set from the arterial pressure waveform data comprising (m) and a parameter (n) based on the patient's body surface area;
The method of claim 1, wherein the method is determined by applying the second multivariate statistical model using evaluating the approximation function using the arterial pressure parameter set. The patient's cardiovascular parameters are:
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on arterial pressure average value in the arterial pressure waveform data set, pressure in the arterial pressure waveform data set A parameter (g) based on a standard deviation of weights, a parameter (h) based on an average value of pressure weights in the arterial pressure waveform data set, a parameter (i) based on the skewness of arterial pulsations in the arterial pressure waveform data set, Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform data Parameter (k) based on pressure-weighted skewness in the set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set (M) an approximation function that is a function of the parameter (n) based on the body surface area of the patient, the approximation relating to a clinically generated reference measurement set representing a blood pressure parameter according to the cardiovascular parameter Determining a function and a clinically determined reference measurement set representing blood pressure parameters in response to the cardiovascular parameters from a patient not experiencing an abnormal condition;
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on the arterial pressure average value in the arterial pressure waveform data set, Parameter (g) based on standard deviation of pressure weight, parameter (h) based on average value of pressure weight in arterial pressure waveform data set, parameter (i) based on arterial pulsation skewness in arterial pressure waveform data set Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform Parameter (k) based on pressure-weighted skewness in the data set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set Determining an arterial pressure parameter set from the arterial pressure waveform data comprising (m) and a parameter (n) based on the patient's body surface area;
The method of claim 1, wherein the method is determined by applying the third multivariate statistical model created using evaluating the approximation function using the arterial pressure parameter set.   The method of claim 1, wherein the abnormal condition represents the occurrence of vasodilation.   The method of claim 1, wherein the abnormal condition represents the occurrence of vasoconstriction.   The method of claim 1, wherein the abnormal condition represents the occurrence of a cardiovascular condition at high cardiac output.   The method of claim 1, wherein the abnormal condition represents a separation in peripheral cardiac pressure high cardiac output from central aortic pressure.   The method of claim 1, wherein the abnormal condition indicates that peripheral arterial pressure is less than central aortic pressure.   The method of claim 1, wherein the abnormal condition indicates that peripheral arterial pressure is not proportional to central aortic pressure. A method for measuring cardiovascular parameters of patients with high cardiac output and non-high cardiac output comprising:
Providing patient arterial pressure waveform data;
Analyzing the arterial pressure waveform data to determine if the patient's peripheral arterial pressure has been separated from the central aortic pressure;
If it is determined that the patient's peripheral arterial pressure has been separated from the central aortic pressure, analyzing the arterial pressure waveform data to determine whether the patient has high cardiac output;
If the patient is determined to have high cardiac output, a third multivariate statistical model is applied to the arterial pressure waveform data to provide the patient's high cardiac output and separated cardiovascular parameters And determining
If the patient's peripheral arterial pressure is not determined to be separated from the central aortic pressure, or the patient is not determined to have high cardiac output, a fourth multivariate statistical model is applied to the arterial pressure waveform data And determining normal cardiovascular parameters for said patient. A method for measuring cardiovascular parameters of patients with high cardiac output and non-high cardiac output comprising:
Providing patient arterial pressure waveform data;
Analyzing the arterial pressure waveform data to determine if the patient's peripheral arterial pressure has been separated from the central aortic pressure;
If it is determined that the patient's peripheral arterial pressure has been separated from the central aortic pressure, analyzing the arterial pressure waveform data to determine whether the patient has high cardiac output;
From an arterial pressure waveform data set from a group of examined patients who were experiencing high cardiac output status and separation between central and peripheral arterial pressure when the patient was determined to have high cardiac output A third multivariate statistical model is created and applied to the arterial pressure waveform data to provide a high cardiac output for the patient and a value for the isolated cardiovascular parameter to apply the high cardiac output for the patient. Determining vascular parameters;
If the patient's peripheral arterial pressure is not determined to be separate from the patient's central aortic pressure, or if the patient is not determined to have high cardiac output, the test patient in a normal hemodynamic state A fourth multivariate statistical model created from the arterial pressure waveform data set from the group and providing the value of the patient's normal cardiovascular parameters is applied to the arterial pressure waveform data to apply the patient's normal cardiovascular Determining a parameter. A method for measuring cardiovascular parameters of patients with high cardiac output and non-high cardiac output comprising:
Providing patient arterial pressure waveform data;
Experienced a first arterial pressure waveform data set from a first group of test patients who had experienced a separation between peripheral arterial pressure and central aortic pressure, and a separation between peripheral arterial pressure and central aortic pressure A first multivariate statistical model created from a second group of arterial pressure waveform data sets from a second group of test patients, the arterial pressure of the first arterial pressure waveform data set Providing a separated output value corresponding to a first separation value in a waveform and a second separation value in the arterial pressure waveform of the second arterial pressure waveform data set, wherein the separated output value is the A first multivariate statistical model that determines that the patient's peripheral arterial pressure has been separated from the central aortic pressure if greater than a separation threshold between a first separation value and the second separation value; Applying to the arterial pressure waveform data the peripheral arterial pressure of the patient And determining whether or not separated from the central aortic pressure,
A first arterial pressure waveform data set from a first group of patients having a high cardiac output and a high cardiac output when the patient's peripheral arterial pressure is determined to be separated from the central aortic pressure; A second multivariate statistical model created from a second arterial pressure waveform data set from a second group of patients that was not a quantity, wherein the arterial pressure waveform of the first arterial pressure waveform data set Providing an output value of a high cardiac output corresponding to a first high cardiac output value at and a second high cardiac output value in the arterial pressure waveform of the second arterial pressure waveform data set And an output value of the high cardiac output is equal to or higher than a threshold value of the high cardiac output between the specified value of the first high cardiac output and the specified value of the second high cardiac output. A second multivariate statistical model that determines that the patient has high cardiac output is applied to the arterial pressure waveform data, And determining whether the cardiac output,
Arterial pressure waveform data set from a group of examined patients who were experiencing high cardiac output status and separation between central and peripheral arterial pressure when said patient was determined to have high cardiac output And applying a third multivariate statistical model to the arterial pressure waveform data to provide a high cardiac output of the patient and the value of the isolated cardiovascular parameter applied to the arterial pressure waveform data. And determining isolated cardiovascular parameters;
If the patient's peripheral arterial pressure is not determined to be separated from the patient's central aortic pressure, or if the patient is not determined to have high cardiac output, from a group of test patients in normal hemodynamic state A fourth multivariate statistical model created from the arterial pressure waveform data set and providing the value of the patient's normal cardiovascular parameter is applied to the arterial pressure waveform data to determine the patient's normal cardiovascular parameter A method comprising:   19. A method according to any one of claims 16, 17 or 18, wherein the high cardiac output and isolated cardiovascular parameters represent the occurrence of vasodilation.   19. A method according to any one of claims 16, 17 or 18, wherein the high cardiac output and isolated cardiovascular parameters represent the occurrence of vasoconstriction.   19. The method of any one of claims 16, 17 or 18, wherein the high cardiac output and isolated cardiovascular parameters represent the occurrence of a high cardiac output cardiovascular condition.   19. The high cardiac output and the separated cardiovascular parameter represent a separation in peripheral cardiac pressure high cardiac output from central aortic pressure, 19. the method of.   19. A method according to any one of claims 16, 17 or 18, wherein the high cardiac output and isolated cardiovascular parameters represent peripheral arterial pressure is less than central aortic pressure.   19. The method according to any one of claims 16, 17 or 18, wherein the high cardiac output and isolated cardiovascular parameters represent peripheral arterial pressure not proportional to central aortic pressure.   Determining whether the patient's peripheral arterial pressure has been separated from the central aortic pressure is a first from a first group of test patients who have experienced a separation between the peripheral arterial pressure and the central aortic pressure. A first arterial pressure waveform data set and a second arterial pressure waveform data set from a second group of test patients who have not experienced a separation between peripheral arterial pressure and central aortic pressure. A first value in the arterial pressure waveform of the first arterial pressure waveform data set, and a second value in the arterial pressure waveform of the second arterial pressure waveform data set, The patient's peripheral arterial pressure is central if the separated output value is greater than a separation threshold between the first value and the second value. A first multivariate statistical model that determines that it has been separated from aortic pressure The includes determining whether peripheral arterial pressure of the patient is applied to the arterial pressure waveform data are separated from the central aortic pressure A method according to any one of claims 16 or 17.   26. The method of claim 25, wherein the first multivariate statistical model is based on a set of elements including one or more parameters that are affected by separation of peripheral arteries.   Determining whether the patient has high cardiac output includes determining a first arterial pressure waveform data set from a first group of patients that had high cardiac output, and high cardiac output. A second multivariate statistical model created from a second arterial pressure waveform data set from a second group of patients not in the arterial pressure waveform of the first arterial pressure waveform data set Providing an output value of high cardiac output corresponding to a first value and a second value in the arterial pressure waveform of the second arterial pressure waveform data set; A second multivariate statistical model that determines that the patient has a high cardiac output if is greater than or equal to a high cardiac output threshold between the first value and the second value; The method according to claim 16, comprising applying to the arterial pressure waveform data.   28. The method of claim 27, wherein the second multivariate statistical model is determined based on an element set that includes one or more parameters that are affected by the high cardiac output status.   26. A method according to any one of claims 2, 18 or 25, wherein the first value is greater than the second value.   26. A method according to any one of claims 2, 18 or 25, wherein the first value is a positive number and the second value is a negative number.   26. A method according to any one of claims 2, 18 or 25, wherein the first value is +100 and the second value is -100.   26. A method according to any one of claims 2, 18 or 25, wherein the threshold is zero.   26. A method according to any one of claims 2, 18 or 25, wherein the separation threshold is an average value of the first separation value and the second separation value.   The separation threshold is a separation threshold range, and if the separated output value is within the separation threshold range, the result is indeterminate and the additional arterial pressure waveform data from the patient is used to 26. A method according to any one of claims 2, 18 or 25, wherein the method is repeated.   The third multivariate statistical model is generated from an arterial pressure waveform data set from a group of examined patients who have experienced high cardiac output status and separation between central and peripheral arterial pressure, The method of claim 16, wherein the method provides cardiac output and separated cardiovascular parameter values.   17. The fourth multivariate statistical model is generated from an arterial pressure waveform data set from a group of test patients in normal hemodynamic state and provides values for the patient's normal cardiovascular parameters. the method of.   37. The method of claim 35 or 36, wherein the multivariate statistical model is based on a set of elements that includes one or more parameters that are affected by the cardiovascular parameters.   The one or more parameters include a parameter (a) based on a standard deviation of pulsation in the arterial pressure waveform data set, a parameter (b) based on an R-to-R interval in the arterial pressure waveform data set, and the arterial pressure waveform data. A parameter (c) based on a region under a cardiac contraction in the set, a parameter (d) based on a cardiac contraction period in the arterial pressure waveform data set, a parameter (e) based on a cardiac dilation period in the arterial pressure waveform data set, the artery Parameter (f) based on arterial pressure average value in pressure waveform data set, parameter (g) based on standard deviation of pressure weight in arterial pressure waveform data set, parameter based on average value of pressure weight in arterial pressure waveform data set (H) Arterial pulsation distortion in the arterial pressure waveform data set (I), a parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, a parameter (k) based on pressure-weighted skewness in the arterial pressure waveform data set, and the arterial pressure waveform 27. The parameter (l) based on pressure-weighted kurtosis in the data set and selected from the group consisting of parameters (m) based on pressure-dependent wind Kessel compliance in the arterial pressure waveform data set. , 28 or 37. The patient's high cardiac output cardiovascular parameters are:
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on the arterial pressure average value in the arterial pressure waveform data set, Parameter (g) based on standard deviation of pressure weight, parameter (h) based on average value of pressure weight in arterial pressure waveform data set, parameter (i) based on arterial pulsation skewness in arterial pressure waveform data set Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform Parameter (k) based on pressure-weighted skewness in the data set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set (M), and an approximate function that is a function of the parameter (n) based on the body surface area of the patient, relating to a clinically generated reference measurement set representing blood pressure parameters as a function of the cardiovascular parameters Determining an approximation function and a clinically determined reference measurement set representing blood pressure parameters in response to the cardiovascular parameters from a patient experiencing a high cardiac output condition;
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on the arterial pressure average value in the arterial pressure waveform data set, Parameter (g) based on standard deviation of pressure weight, parameter (h) based on average value of pressure weight in arterial pressure waveform data set, parameter (i) based on arterial pulsation skewness in arterial pressure waveform data set Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform Parameter (k) based on pressure-weighted skewness in the data set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set Determining an arterial pressure parameter set from the arterial pressure waveform data comprising (m) and a parameter (n) based on the patient's body surface area;
Calculating the high cardiac output cardiovascular parameter of the patient by evaluating the approximation function using the arterial pressure parameter set, and applying the third multivariate statistical model 19. A method according to any one of claims 16, 17 or 18, wherein: The patient's normal cardiovascular parameters are:
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on the arterial pressure average value in the arterial pressure waveform data set, Parameter (g) based on standard deviation of pressure weight, parameter (h) based on average value of pressure weight in arterial pressure waveform data set, parameter (i) based on arterial pulsation skewness in arterial pressure waveform data set Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform Parameter (k) based on pressure-weighted skewness in the data set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set (M), and an approximate function that is a function of the parameter (n) based on the body surface area of the patient, relating to a clinically generated reference measurement set representing blood pressure parameters as a function of the cardiovascular parameters Determining an approximation function and a clinically determined reference measurement set representing blood pressure parameters in response to the cardiovascular parameters from a patient in a normal hemodynamic state;
At least the parameter (a) based on the standard deviation of the arterial pressure waveform data, the parameter (b) based on the heart rate of the patient, the parameter (c) based on the region under the cardiac contraction part of the arterial pressure signal, and the cardiac contraction period Parameter (d), parameter (e) based on the ratio between the systole period and the diastole period, parameter (f) based on the arterial pressure average value in the arterial pressure waveform data set, Parameter (g) based on standard deviation of pressure weight, parameter (h) based on average value of pressure weight in arterial pressure waveform data set, parameter (i) based on arterial pulsation skewness in arterial pressure waveform data set Parameter (j) based on kurtosis of arterial pulsation in the arterial pressure waveform data set, the arterial pressure waveform Parameter (k) based on pressure-weighted skewness in the data set, parameter (l) based on pressure-weighted kurtosis in the arterial pressure waveform data set, parameter based on pressure-dependent wind kessel compliance in the arterial pressure waveform data set Determining an arterial pressure parameter set from the arterial pressure waveform data comprising (m) and a parameter (n) based on the patient's body surface area;
Calculating the cardiovascular parameters in the normal hemodynamics of the patient by evaluating the approximation function using the arterial pressure parameter set, and applying the fourth multivariate statistical model 19. A method according to any one of claims 16, 17 or 18, wherein:   41. The method of any one of claims 1, 16, 17, 18, 39, or 40, wherein the cardiovascular parameter is arterial compliance.   41. The method of any one of claims 1, 16, 17, 18, 39, or 40, wherein the cardiovascular parameter is arterial elasticity.   41. The method of any one of claims 1, 16, 17, 18, 39, or 40, wherein the cardiovascular parameter is peripheral resistance.   41. The method of any one of claims 1, 16, 17, 18, 39, or 40, wherein the cardiovascular parameter is arterial tone.   41. The method of any one of claims 1, 16, 17, 18, 39, or 40, wherein the cardiovascular parameter is arterial flow.   41. The method of any one of claims 1, 16, 17, 18, 39, or 40, wherein the cardiovascular parameter is stroke volume.   41. The method of any one of claims 1, 16, 17, 18, 39, or 40, wherein the cardiovascular parameter is cardiac output.   19. The method of any one of claims 1, 16, 17, or 18, wherein the arterial pressure waveform data from the patient is provided continuously and analyzed continuously.

Images